Optimal Chunk Scheduling Algorithm Based On Taboo Search For Adaptive Live Video Streaming In CDN-P2P

With a massive increase in user device heterogeneity, an effective distribution of live video streaming that takes into consideration varying display sizes and processing capabilities of end user devices is becoming a necessity for live multimedia content providers worldwide. In order to answer the challenges of device heterogeneity, an underlying delivery architecture needs to be enhanced so that the best possible quality of video is delivered to the end user in accordance with the end users' device capabilities in a timely and effective fashion. In a primitive video streaming environment that does not take device heterogeneity into consideration, video streaming is based on a single-layered video which uses a scheduling algorithm that ensures timely download of video chunks (i.e. chunks are downloaded before they need to be played back). This paper introduces a model for chunk scheduling algorithm which examines the issue of live video streaming performance optimization as a knapsack NP-hard problem. The proposed scheduling algorithm uses a taboo search method to identify and prioritize chunk delivery in addition to taking bandwidth throughput and delivery ratio into consideration. It optimizes bandwidth utilization by identification of "useless available chunks" by identifying them via chunk playback deadline. With this approach, chunks whose playback deadline has passed as well as those who belong to layers that do not have corresponding lower values will be considered as useless and will be disposed of. Finally, this paper provides simulation results which show that the proposed algorithm significantly performs better than the 'primitive' chunk scheduling strategies especially in situations with rigid bandwidth constraints.